Coupled Split-Ring Resonators for Isolation Improvement in a 1 × 2 Microstrip Patch Antenna Array

Abstract

In this paper, a method to reduce mutual coupling between an E-plane and H-plane coupled microstrip patch antenna is presented. Two dual differentially fed square patches are designed in a 1 × 2 antenna array configuration. To minimize mutual coupling and its effects, coupled split-ring resonators (SRRs) are designed, characterized and positioned between the patches. Circular SRRs are designed and coupled to produce a band-stop response to suppress surface waves propagating within the dielectric substrate while enhancing isolation. Mutual coupling interactions and the suppression mechanism are discussed in relation to the patches and SRRs. The patch radiators are dual differentially fed to achieve polarization diversity. E- and H-planes decoupling is achieved between the two patches throughout their bandwidth while maintaining good antenna performance. A prototype of the antenna array and the SRR is fabricated and measured to validate the decoupling approach. With a separation distance of 0.49$\lambda$ between the patches, the measured S-parameters show an impedance bandwidth of |S11|≤−10 dB, covering 9.27–9.46 GHz, and −38 dB and −35 dB mutual coupling for E- and H-planes, respectively, are observed throughout the antenna operating bandwidth.

1. Introduction

Nowadays, the rising demand for high-performance and compact antennas have prompted the development of array antennas for many wireless communication applications [1]. Patch antennas are famous and widely applicable in antenna array designs owing to their advantages of compactness, planar geometry and ease of fabrication and integration [1]. Nonetheless, when multiple patch radiating elements are placed close to each other on the same substrate to achieve improved performance such as beamforming, high gain and polarization diversity, significant coupling occurs due to surface and near-field space waves [2,3]. The mutual coupling degrades the overall antenna performance characteristics in terms of impedance matching, resonance frequency shift and radiation pattern [1,4,5]. Improving isolation between antenna elements is often considered in designing tightly packed antenna arrays. Consequently, multiple approaches have been developed to tackle mutual coupling and improve isolation between patches in a two-element array configuration [4,6]. The employment of defected ground structures (DGSs) is a proven technique to minimize inter-element coupling considering that most times it does not involve the addition of extra elements but suffers from high back radiation [7,8,9]. Neutralization lines between two antennas are shown to also be effective in reducing mutual coupling [10]. However, as stated in [7] the main drawback of the neutralization line is its narrow bandwidth characteristics. The application of metamaterials and other resonant structures is also an efficient technique in mitigating mutual coupling [11,12,13,14,15,16]. In [2,17], the utilization of split-ring resonators was demonstrated to achieve mutual coupling suppression, particularly in one of the principal planes (E- or H-plane) of the antenna. This paper proposes an approach to simultaneously decouple an E- and H-plane coupled microstrip patch antenna with coupled SRRs. The proposed design is a modification of [18], where the SRR is removed from the ground plane to minimize back radiation common in defected ground structures; it is fabricated to validate the proposed antenna with the SRR. Through numerical and experimental investigations, the proposed design achieved a |S11|≤−10 dB bandwidth of 9.27–9.46 GHz and a measured maximum broadside gain of 6.8 dBi, while having isolation improvements of up to 12 dB in the E-plane and 6 dB in the H-plane, with an inter-element spacing of 0.49$\lambda$.

2. Antenna Structure and Design Considerations

The proposed antenna configuration incorporating SRRs is shown in Figure 1. Patch radiators and SRRs are designed and implemented on the topside of substrate 1. The backside of substrate 2 is the feeding network, while its topside is the ground plane. Both substrates stack together and they are Rogers RO5880 (${ϵ}_r$ = 2.2, tan$\delta$ = 0.0009 and $h_1$ = $h_2$ = 0.508 mm). The SRRs are separated by a spacing of 0.39 mm and each SRR connected to the ground plane with vias. Each patch is excited with four feeding-through hole vias attached to the differential feed lines. In Figure 1, pairs 1 and 2 on each patch are orthogonal and connected to a 180° delay line or a 0.5$\lambda$ microstrip line [19,20]. The antenna was optimized with the following design parameters: F_w = 1.37 mm, $W_p$ = 10.16 mm, $F_t$ = 0.38 mm, $R_r$ = 3.54 mm, L = 62 mm, $W_r$ = 0.79 mm, W = 36 mm, $g_1$ = 0.8 mm and $F_l$ = 4.88 mm.

3. SRR Characterization and Antenna Performance Analysis

To mitigate mutual coupling, two magnetically coupled resonators with a band-stop filter response were placed between the patches. The layout of the coupled SRR is shown in Figure 2. The decoupling structure is made of SRRs comprising printed metal rings etched on a dielectric substrate of thickness 1.6 mm with splits on their opposing sides. A circular geometry was considered since rings are simple to design and implement with less design parameters. Fundamentally, SRRs at high frequencies are resonant structures exhibiting high quality [21]. When illuminated through a changing incoming magnetic field that is parallel to the axis of the SRR, an electromotive force is generated along the ring that further produces current loops, and the distributed capacitance of the ring also serves to close the current loops [22,23]. Therefore, the SRR responds as a resonant LC circuit that is externally driven and can be tuned through the SRR dimensions of ring radius, split width and ring width [23]. Details on resonant frequency calculations and design formulas are well explained in [24].

In addition, as demonstrated in [21,23,25], a periodic structure or array of split-ring resonators are excited by an incoming magnetic field parallel to the ring axis, with the electric field along the split width; strong magnetic dipole moments are excited at resonance and the SRR is capable of inhibiting wave propagation, particularly around its resonance frequency [23,26]. This can be attributed to the highly negative effective permeability within a small frequency range around the resonant frequency, which is a property of periodic media [23]. The reflection or rejection level is dependent on the number of SRRs, coupling degree and the distance between adjacent SRRs. To characterize the decoupling SRR, a two-port network electromagnetic model is set up and the SRR is excited with a plane wave across the coupled SRR, and the reflection and transmission characteristics can be studied using S-parameters. The setup and analysis of the model are detailed in [27]. Figure 2 demonstrates the |S11| and |S21| characteristics of the coupled SRR. The transmission profile reveals significant attenuation within the 9.37 GHz frequency range. The figure demonstrates that the coupled SRR generates a band rejection response within the designed frequency band of the microstrip patch.

The predominant mechanism of mutual coupling is via surface waves propagating within the substrate. Due to the orientation of differential pairs 1 and 2, which are orthogonal to each other, the patches are coupled simultaneously in both E- and H-planes. The patches are coupled in the E-plane via surface waves and are slightly coupled in the H-plane via space waves. The coupled SRRs decouple the patch in the E–plane by suppressing surface waves propagating within the dielectric substrate with its band-stop response characteristics. The via generates extra inductance to the ring that shifts its resonance frequency to the desired operating band and increases the sharpness of the resonance of the resonator by increasing its quality factor. In addition, the SRR in conjunction with the vias provides an alternative coupling part for space waves, improving the isolation between the two patches in the H-plane. The effectiveness of the method in enhancing isolation is evaluated with the SRR in a desired position between the patches. In reference to Figure 1, the coupling from patch A to B and vice versa is characterized by $|S{1}_B{1}_A|$ and $|S{2}_B{2}_A|$. The S-parameter $|S{1}_B{1}_A|$ is the H-plane coupling and it is interpreted as coupling from port 1 on patch A (differential pair 1A) to port 1 on patch B (differential pair 1B), while $|S{2}_B{2}_A|$ represents E-plane coupling from port 2 on patch A (differential pair 2A) to port 2 on patch B (differential pair 2B). The mutual coupling profile of the antenna array with and without the coupled SRR is expressed in Figure 3 and Figure 4. Figure 3 compares the isolation between 1A and 1B while Figure 4 demonstrates the isolation between 2A and 2B with and without the coupled SRR. The placement of the coupled SRR between the two patches significantly attenuates the propagation of surface waves, effectively enhancing isolation. This decoupling mechanism resulted in additional isolation enhancements of the 12 dB E-plane and 6 dB H-plane, as confirmed by measurements.

The reduction in surface waves through the band rejection characteristic of the coupled SRR is clearly observed from the current distribution on the patches shown in Figure 5. A high current concentration is observed on the unexcited patch when the decoupling SRR is absent. However, when the coupled SRR is placed between the patches, the surface waves generated by the excited patch are attenuated, leading to a lower surface current concentration on the unexcited patch.

4. Experimental Validation

A prototype of the antenna incorporated with coupled SRR was fabricated and measured. The patches were excited through SMPM connectors attached to the differential feeding networks as shown in Figure 6. To measure the scattering parameters, Rohde and Schwarz ZVA 67 VNA was used. As shown in Figure 7, all antenna ports measured a |S11|≤−10 dB bandwidth of 9.27–9.46 GHz. The mutual coupling between the patches was measured and the results are presented in Figure 8 and Figure 9. Mutual coupling measurements reveal a significant increase in isolation due to the introduction of the coupled SRR. The slight differences in the profile of the measured and simulated mutual coupling can be attributed to extra inductance introduced by the cables during the measurement setup.

The antenna array demonstrates strong performance characteristics. Simulated reflection coefficients for all four ports indicate resonance at 9.37 GHz, with good impedance matching and an impedance bandwidth ranging from 9.27 to 9.46 GHz. As shown in Figure 10, the antenna maintains stable gain and exhibits a fairly symmetric unidirectional radiation pattern across its operating bandwidth. The integration of the coupled SRR does not significantly degrade the radiation pattern and achieves a favorable cross-polarization ratio. The polarization diversity capabilities of the antenna are thoroughly discussed in [19,20].

Table 1. Performance comparison of the proposed antenna.

Ref.	Method	Freq (GHz)	BW (GHz)	Coupling Plane	Max. Isolation Improvement (dB)
[1]	U-shaped resonator	5.0	4.95–5.15	E	14
[11]	SCCSRR	3.7	3.65–3.75	E	19
[13]	Slotted meander line resonator	4.8	4.75–4.95	H	16
[14]	Meander line resonator	2.8	2.70–2.80	H	10
[15]	Decoupling resonator	3.5	3.47–5.52	H	16
[16]	Planar resonator	2.93	2.77–3.10	-	7
[17]	Slotted CSRR	5.0	4.95–5.05	E	10
Prop.	Coupled SRR	9.36	9.27–9.46	E/H	12/6

BW: Bandwidth. Freq: Center frequency.

presents a performance comparison of the proposed design and other resonator-based structures used for mutual coupling suppression. Notably, the coupled SRR effectively suppresses E- and H-plane couplings, addressing the dual differential feeding of the radiating patches. Though the design produces slightly lower isolation improvements when compared with other resonator-based decoupling solutions, the performance of the antenna design and the coupled SRR is still satisfactory and falls within acceptable limits for many applications while being unique in providing simultaneous decoupling in both E- and H-planes.

Antenna far-field characteristics were measured in an anechoic chamber with a standard horn antenna MI-12A-26. The gain at different frequencies was measured and is shown in Figure 10. The measured gain is almost identical to the simulated gain profile with a measured maximum gain of 6.8 dBi. Differences in the simulated and measured gain can be attributed to insertion loss introduced by the cables and power dividers in the pattern measurement setup. The antenna maintains stable gain and exhibits a fairly symmetric unidirectional radiation pattern as seen in Figure 11 across its operating bandwidth. The integration of the coupled SRR does not significantly degrade the radiation pattern and achieves a favorable ratio of co-polarization to cross-polarization.

5. Conclusions

This paper proposed the implementation of a coupled SRR for mutual coupling reduction between two dual differentially fed microstrip patches. The coupled SRR was simulated to characterize and demonstrate band-stop properties from its scattering parameters. The introduction of the coupled SRR between the patches resulted in significant mutual coupling reduction. The method shows that mutual coupling reduction and isolation improvements of 12 dB and 6 dB in the E- and H-planes, respectively, are attainable. The approach has potentials of being extended to larger arrays, which is under consideration for future work. A fabricated sample with measurement results validates the coupled SRR as a suitable technique for decoupling the closely spaced antenna arrays. With a measured 10 dB impedance bandwidth of 9.27–9.46 GHz, a broadside measured maximum gain of 6.8 dBi and a fairly symmetric radiation pattern, the antenna demonstrates suitability for polarization diversity applications.